A Number Above the Threshold, an Exam That Doesn't Match It
The guideline threshold says insulin. Her exam says something less urgent. The real disagreement isn't about the number itself — it's about which signal should actually drive today's decision.
Nadia F., a 39-year-old woman, manages the books for her family's auto repair shop and put off her annual physical for nearly two years, mostly out of a busy schedule rather than avoidance. When she finally came in, a routine screen turned up a fasting glucose of 268 mg/dL, confirmed on repeat, with an A1c of 10.8% and no prior diagnosis. What her exam and history don't show is anything catabolic: her weight has been stable for the past year, she denies any unusual thirst or urination she'd specifically noticed, and her urine ketones are negative. She's frightened by the number itself and has told the team directly she'd strongly prefer to avoid starting insulin if there's a reasonable alternative.
Her A1c alone puts her above the threshold the ADA Standards of Care uses to recommend starting insulin at diagnosis, on the reasoning that insulin most reliably reverses glucotoxicity — the state in which very high glucose itself further impairs the body's remaining insulin-secreting capacity, creating a cycle that can be hard for oral or non-insulin injectable therapy to break on its own at this severity. But her clinical picture doesn't show the catabolic features — significant weight loss, ketosis, dehydration — that typically accompany the most severe, insulin-deficient presentations the guideline threshold was built to catch. The threshold was never really about the A1c itself; it was built as a stand-in for insulin deficiency severe enough to have started consuming the patient. In her the proxy and the thing it stands for have come apart, and a threshold cannot arbitrate the one case it was designed to assume away.
A number above the threshold, an exam below it
I'd start insulin, at least initially. An A1c above 10% is where multiple guidelines recommend insulin as initial therapy, precisely because it most reliably reverses glucotoxicity. Starting a sub-maximal combination and hoping it's enough risks weeks of continued severe hyperglycemia if it isn't — insulin gets her to safety fastest, and we can de-escalate once she's there.
I'd weigh her actual exam more heavily than the isolated number. She has none of the catabolic features — weight loss, ketosis, real dehydration — that typically accompany the most severe, insulin-deficient presentations the A1c threshold was meant to catch. A maximal combination of metformin, an SGLT2 inhibitor, and a GLP-1 receptor agonist together can achieve real, substantial glycemic improvement without insulin's hypoglycemia risk or the injection burden she's told us directly she wants to avoid.
I'm not dismissing the A1c — it's genuinely high. I'm saying the absence of catabolic symptoms is a real, clinically meaningful signal too, and treating the number alone as decisive ignores information her exam is giving us.
I think there's a way to get what both of you actually want. Start insulin now, but frame it explicitly as a time-limited bridge, not a permanent regimen — get her glucose down reliably and quickly, then plan de-escalation to a non-insulin combination once glucotoxicity resolves and we can see her true beta-cell function.
That's not a compromise for its own sake. It captures insulin's real reliability at reversing this severity of hyperglycemia while directly respecting her stated preference, framed honestly as temporary rather than open-ended — which also makes it a much easier conversation to have with her today.
Basal insulin started alongside metformin, with a documented plan to reassess at eight to twelve weeks and begin transitioning to a non-insulin combination regimen once her glucose has stabilized. Nadia left the visit understanding insulin as a defined bridge rather than a permanent commitment, which the team noted changed how she received the recommendation.
Not agreed: whether eight or twelve weeks is the right interval for the de-escalation reassessment. The endocrinologist preferred the full twelve weeks to ensure genuine glucotoxicity resolution before judging her true beta-cell function; the primary care physician wanted an eight-week check given her strong preference to come off insulin as soon as reasonably possible. Left as a range, to be finalized based on how quickly her glucose actually normalizes.